Research On Crowd-sensing Trust Management Mechanism In Crowd Evacuation

With the advancement of modernisation,people’s quality of life is becoming higher and higher,and there are more and more large gatherings.However,in recent years,accidents have been frequent at these gatherings,and the lack of proper evacuation guidance can lead to panic in the crowd,resulting in stampede accidents and even major casualties.The use of computer simulation technology to simulate the evacuation process of crowd in a dense place can not only identify a reasonable response strategy,but also save the cost of on-site simulation.As a result,computer simulation has received increasing attention from researchers in the field of crowd evacuation simulation.Crowd evacuation is a process that is influenced by many factors.Individual trust is a very important factor in crowd evacuation.Psychological research has found that individual trust has a direct impact on the psychological state and emotions of individuals,which in turn affects their movement decisions.In crowd evacuation,people with high levels of trust play a very important role in guiding the evacuation and planning the route.However,existing crowd evacuation methods do not effectively manage the trust level of individuals in the crowd,making it impossible to accurately guide evacuation using high-trust personnel in a timely manner,resulting in a lack of realism and inefficiency in the evacuation simulation.Therefore,it is a challenging problem to find a reasonable evacuation response strategy for crowd evacuation and how to effectively manage the trust of individuals in the crowd.To solve the above issues,this thesis proposes a crowd-sensing trust management mechanism in crowd evacuation,which manages the trust level of individuals during evacuation and selects people with high trust level to guide the crowd evacuation.Firstly,a blockchain based crowd sensing trust management mechanism(B-CSTM)is proposed to manage the individual trust level in the crowd.Secondly,based on the above trust management mechanism,a trust-based crowd evacuation path planning method is proposed,a trust-based probabilistic roadmap(T-PRM)algorithm is constructed to obtain the trusted paths,and the trusted paths are combined with the reciprocal velocity obstacles(RVO)model to obtain the crowd collision-free motion trajectory.Finally,based on the above methods,a trust-based crowd evacuation simulation system is constructed to simulate the crowd evacuation process.The main work and innovation of this thesis are as follows:(1)To solve the problem that current crowd evacuation studies do not effectively manage crowd trust,this thesis proposes a blockchain-based crowd sensing trust management mechanism.First,a message evaluation mechanism is proposed,which uses crowd sensing to sense messages and judge the trustworthiness of the messages by evaluating them.Secondly,the evacuation perception unit(EPU)is defined to calculate the trustworthiness of individuals,and message trustworthiness and person history trustworthiness are used as two-dimensional feature variables to calculate person’s trustworthiness using a Bayesian model.Finally,blockchain is used to manage person trustworthiness,and a hybrid consensus algorithm combining proof of stake(POS)and practical byzantine fault tolerance(PBFT)is proposed to improve the consensus efficiency of the block.The experiments demonstrate that this mechanism can accurately capture the individual trust level in the evacuation process and use the blockchain for decentralized trust management,improving the security of trust management and maintaining data consistency.(2)To solve the problem that the guiding role of high trustworthy personnel is not fully considered in crowd evacuation path planning,this thesis proposes a trust-based crowd evacuation path planning method.At the macro level,the T-PRM algorithm is proposed to construct a path navigation graph with trust cost and search for trusted paths;at the micro level,the T-PRM-based crowd motion model is proposed to obtain the crowd collision-free motion trajectory by combining the trusted paths with the RVO model.Experiments prove that the method can reasonably and efficiently complete path planning in complex scenes.(3)A trust-based crowd evacuation simulation system is constructed.The crowd movement based on the method in this thesis is rendered in a 3D platform for output.The analysis of the realistic rendering results proves that the method is feasible and realistic.